Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/artttj/llama-steve/agents-mdgit clone --depth 1 https://github.com/artttj/llama-steveWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/artttj/llama-steve/agents-md)<a href="https://agentmods.dev/instructions/artttj/llama-steve/agents-md"><img src="https://agentmods.dev/badge/instructions/artttj/llama-steve/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00817 | $0.00817 |
| Opus 5 | $0.00409 | $0.00409 |
| Sonnet 5 | $0.00163 | $0.00163 |
| Haiku 4.5 | $0.00082 | $0.00082 |
Grade A, and why
llama-steve AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Llama Steve
Guidance for coding agents (Claude Code, etc.) working in this repo.
What this is
Llama Steve is a Claude Code plugin that runs a 10x product teardown of a repo
in the voice of a Steve-Jobs-modeled persona — it generates game-changing ideas,
says no to most of them, keeps ~3, and stamps a binary verdict (SHEEP IT or
SPIT IT). It renders the result as an Apple-keynote-style report and serves it
on localhost. A curated landing page is published at https://sonto.space/steve/.
Architecture (read before changing anything)
The CLI is a deterministic renderer. It makes zero LLM calls. The actual
thinking is done by the Claude Code agent following commands/llama-steve.md.
llama-steve.mjs— CLI. Verbs:serve(render/tmp/llama-steve-result.jsonintoreports/and serve it) andlanding(renderassets/demo.jsonintosite/for the public demo). Flags:--open,--no-serve,--port,--result.lib/report.mjs—renderReport/buildReport. Emits self-contained HTML and copiesassets/. Two modes:report(local) andlanding(adds<base href="/steve/">). CSS is cache-busted by content hash.boldNames()lightly bolds known framework/company names; the brand mark isassets/llama-mark.svginlined withfill: currentColor.lib/scan.mjs—repoIdentity(): name/description from manifest, git remote, commit count.lib/serve.mjs— tiny static server, bound to127.0.0.1, path-traversal guarded.skills/— bundled skills used during a teardown:llama-steve(persona + output contract),ten-x(idea generation),humanizer(verdict rewrite).assets/demo.json— data for the public landing. Must be a real teardown.assets/example.jsonis the fallback sample shown when no result exists.
Running it
node llama-steve.mjs serve --open # local report from /tmp/llama-steve-result.json
node llama-steve.mjs landing # build public site/ from assets/demo.json
node --test test/*.mjs # tests (Node 20+, no deps, no build step)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 61 lines · 817 tokens per session scan A 781d42da2fce
llama-steve AGENTS.md is an instructions file published in the GitHub repository artttj/llama-steve (1 stars, last pushed 3mo ago), licensed MIT. It adds 817 tokens to every session, about $0.0041 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
create-claude-plugin CLAUDE.md
Claude Code instructions for codyhxyz/create-claude-plugin, covering create-claude-plugin — claude code plugin, canonical design docs — read first, directory invariants, two-layer validation architecture and path rules.
forge CLAUDE.md
Claude Code instructions for TT-Wang/forge, covering forge, quick start, skills, agents and mcp tools.
markifact-mcp GEMINI.md
Instructions for markifact/markifact-mcp, covering markifact, the 8 tools, discover → inspect → run, connection vs account and reporting workflow.
Version-Sentinel CLAUDE.md
Claude Code instructions for KSEGIT/Version-Sentinel, covering version sentinel, project structure, supported ecosystems, how it works and prerequisites.
magi-workflow CLAUDE.md
Instructions for howar31/magi-workflow, covering what this is, slash commands, subagents, project state model and project document tiers.
Looptimal CLAUDE.md
Instructions for Renn-Labs/Looptimal, covering looptimal — repo guidance for ai agents, build journal — capture the concepts, not the changelog and — .